A5074583465- Functional Brain Connectivity Studies
- Health, Environment, Cognitive Aging
- Statistical Methods and Inference
- Advanced MRI Techniques and Applications
- Machine Learning in Healthcare
- Neural Networks and Applications
- Advanced Neuroimaging Techniques and Applications
- Meta-analysis and systematic reviews
- Mental Health Research Topics
- Adversarial Robustness in Machine Learning
- Genetic Associations and Epidemiology
- Digital Rights Management and Security
- Artificial Intelligence in Healthcare and Education
- Topic Modeling
- Explainable Artificial Intelligence (XAI)
- Statistical Methods and Bayesian Inference
- Imbalanced Data Classification Techniques
- Bone health and osteoporosis research
- EEG and Brain-Computer Interfaces
- Financial Distress and Bankruptcy Prediction
Inria Saclay - Île de France
2022-2024
École Polytechnique
2024
McGill University
2020-2023
Florida International University
2023
Montreal Neurological Institute and Hospital
2020-2022
Institut national de recherche en informatique et en automatique
2022
Mila - Quebec Artificial Intelligence Institute
2022
We present NiMARE (Neuroimaging Meta‑Analysis Research Environment; RRID:SCR_0173981), a Python library for neuroimaging meta‑analyses and metaanalysis‑related analyses. is an open source, collaboratively‑developed package that implements range of meta‑ analytic algorithms, including coordinate‑ image‑based meta‑analyses, automated annotation, functional decoding, meta‑analytic coactivation modeling. By consolidating methods under common syntax, makes it straightforward users to employ the...
Abstract Background As databases grow larger, it becomes harder to fully control their collection, and they frequently come with missing values. These large are well suited train machine learning models, e.g., for forecasting or extract biomarkers in biomedical settings. Such predictive approaches can use discriminative—rather than generative—modeling thus open the door new missing-values strategies. Yet existing empirical evaluations of strategies handle values have focused on inferential...
The ability to ensure that a classifier gives reliable confidence scores is essential informed decision-making. To this end, recent work has focused on miscalibration, i.e., the over or under of model scores. Yet calibration not enough: even perfectly calibrated with best possible accuracy can have are far from true posterior probabilities. This due grouping loss, created by samples same but different Proper scoring rule theory shows given missing piece characterize individual errors loss....
Large Language Models (LLMs), including ChatGPT and LLaMA, are susceptible to generating hallucinated answers in a confident tone. While efforts elicit calibrate confidence scores have proven useful, recent findings show that controlling uncertainty must go beyond calibration: predicted may deviate significantly from the actual posterior probabilities due impact of grouping loss. In this work, we construct new evaluation dataset derived knowledge base assess given Mistral LLaMA. Experiments...
BACKGROUND As databases grow larger, it becomes harder to fully control their collection, and they frequently come with missing values: incomplete observations. These large are well suited train machine-learning models, for instance forecasting or extract biomarkers in biomedical settings. Such predictive approaches can use discriminative --rather than generative-- modeling, thus open the door new missing-values strategies. Yet existing empirical evaluations of strategies handle values have...
BACKGROUND: As databases grow larger, it becomes harder to fully control their collection, and they frequently come with missing values: incomplete observations. These large are well suited train machine-learning models, for instance forecasting or extract biomarkers in biomedical settings. Such predictive approaches can use discriminative -- rather than generative modeling, thus open the door new missing-values strategies. Yet existing empirical evaluations of strategies handle values have...